A least-squares-adjustment-based method for automated map generalization of settlement areas in GIS

Least squares method is one of the effective and widely used data process tool in statistical based data analysis. Automated map generalization is an intelligent process in nature, which involves mining the knowledge from the original data sets to form the generalized features that are the abstracted representation of the original features in multiple and different scale levels. In this paper, we present an integrated methodology for the automated generalization of settlement areas based on least squares adjustment. Taking the original data sets of the features as observational values, the parameters representing the generalized features are computed according to the least squares adjustment model. The four-level hierarchical model for settlement area generalization is first proposed. The rules and constraints conditions in settlement area generalization are then described, and the least squares adjustment model is derived in settlement areas generalization. The results of practical tests demonstrate that the validity and feasibility of the proposed model for the simplification of settlement areas in geographic information system.